A Comment In Response To Michael Pettis On Investor Rationality

In "When are markets 'rational'?", Michael Pettis argues that markets need a number of investor types in order to allocate investment efficiently; in particular, you need value (fundamental) investors to provide some anchor to securities pricing, and speculators to provide liquidity.

Following from the insights of Hyman Minsky (who built upon Keynes and Kalecki), I think it is very difficult to differentiate between "speculation" and "rational investing" in modern capitalist economies. The feedback loop between the financial markets and the real economy means that what may appear to be speculative activity ends up generating the profits that ultimately justifies the "speculators'" investment thesis...



(As an aside, investor rationality seems to be a topic of recent interest, as Michael Pettis notes. As I discuss in this related article, financial markets appear efficient, in the sense that they are hard to beat, but rationality is generally hard to find in the markets for corporate securities. UPDATE: I added a qualification to what I originally wrote here, as rate expectations in the bond market represent a form of market rationality. And that said, the rates markets are unusual in that they have a strong valuation anchor, unlike the markets in corporate securities.)

 In his article, Michael Pettis divides investment strategies into 3 groups:

1. Value, or Fundamental investing, in which securities are bought on the basis of being cheap to long-term fundamentals. They should stabilise the market, by getting out when prices rise too fast, and buying when they fall too far.
2. Relative Value investing, in which securities are valued relative to each other, and no attempt is made to value them in absolute terms. This keeps the overall market pricing coherent.
3. Speculative, or short-term investing. This activity is often trend-following, and injects instability into the system. This instability allows the market to react quickly to new information, and provides liquidity so that value investors can enter or exit positions.

I would add in market-making activity, where I would lump in High Frequency Trading. The environment for market makers influences things like corporate spreads (dealers will not want to hold a lot of inventory if spreads are tighter than their own funding costs).

I am in agreement with Pettis with regards to his comments on the financial market structure. My opinion is that Relative Value investing has probably become too dominant within the investment community as the result of institutional factors. The way performance is measured means that relative value investing is the only way to be a "good" investor, and so most efforts are put into relative value analysis. (Disclaimer: my previous employment was mainly involved with relative value analysis, but I also had a good look at how fundamental analysis was done.) As an extreme example, I would guess that most investors' fair value estimates of USD investment grade 5-year corporate spreads lie within a range of 30-50 basis points. On the other hand, investors' fair value estimates for the level of 10-year U.S. Treasury Note yields lie in a range at least 300-400 basis points wide.

On the other hand, I am skeptical about the ability of Value Investors to stabilise the financial markets or the business cycle. To do so, they would need access to economic models that offer good forecasting ability for economic aggregates like corporate profits. However, for this task, macroeconomic research has probably gone backwards rather than improved over recent decades. Modern macro research is largely dominated by models which only make sense for central bank policy makers - if I move interest rates by 100 basis points, what is the response of inflation? (These models failed to be useful for even that limited area of use during the latest crisis, but that's a topic for another day.)

The underlying problem is that financial  investor behaviour helps determine the real economy outcome. Speculative activity in the financial markets typically leads to higher fixed investment, and thus higher corporate profits. (This follows from the Kalecki profit equation.) As such, it is impossible to give a "fundamental" value for a corporate security without taking into account the impact of the financial markets on economic activity.

It is possible to value securities without taking into account your investing activity essentially in two cases:

1. if you are too small to move the market; or
2. you are analysing a small growth stock that can grow via increasing market share regardless of the direction of the overall economy.

Since there is a great deal of interest in these topics, the lessons learned here end up being incorrectly extrapolated to the aggregate markets.

As Hyman Minsky observed, value investing was dominant in the 1950's, and speculative activity was dormant. This was a special case, and was not a sustainable pattern of activity. The private sector blew itself up spectacularly during the 1930's, and memories of this were fresh. Therefore, speculation in securities markets was not socially acceptable. Meanwhile, private sector balance sheets were loaded up with liquid Treasury securities, making financial markets highly resilient to liquidity-driven panics. This meant that financial markets in the 1950's were relatively tranquil, especially in comparison to the inventory-cycle driven real economy. However, memories of the 1930's faded, and the weight of Treasury holdings dropped, setting the stage for the financial market instability we have observed since then. (This is discussed in Minsky's book, "John Maynard Keynes".)

In conclusion, "modern" financial theory has its roots in the 1960's, and the analysis was mainly back-tested on the tranquil 1950's data (as pre-war data is limited and dominated by the Crash). This has lead to the misleading conclusions about the role of value investing, in my view.

(c) Brian Romanchuk 2013